original article ANALYSIS OF TRENDS IN TOTAL AND AIDS- RELATED DEATHS CERTIFIED AT MOSVOLD HOSPITAL, , KWAZULU-NATAL, FROM 2003 TO 2008

C H Vaughan Williams, MB BS, DCH, MFamMed, DOH District Family Physician, Umkhanyakude Health District Office, , KwaZulu-Natal

Objectives. To analyse mortality trends from deaths registered at Mosvold Hospital, Ingwavuma, KwaZulu-Natal, and possible impact of programmes to treat and prevent HIV infection.

Design. Longitudinal study of death certifications from 2003 to 2008.

Setting. Mosvold Hospital mortuary, Ingwavuma.

Subjects. Counterfoils of form 83/BI-1663, Notification/Register of Death/Stillbirths (Republic of , Department of Home Affairs), completed at Mosvold Hospital from January 2003 to December 2008.

Outcome measures. Age at death, cause of death, patterns of deaths grouped by age, gender and cause of death.

Results. AIDS-related deaths were the cause of 53% of deaths, particularly affecting the 20 - 59-year and under-5 age groups. Since 2005 there has been a decline in deaths in the 20 - 59 age group and an increase in average age at death.

Conclusions. The decrease in mortality from 2005 may be associated with antiretroviral roll-out reducing mortality from AIDS-related illnesses.

Mosvold Hospital is situated in northern KwaZulu- In a previous study,2 an analysis of 4 years’ mortality Natal near the borders of Swaziland and . data from 2003 to 2006 indicated that AIDS-related According to estimates by the Department of Health, illnesses were responsible for 53% of deaths certified the hospital serves a population of about 108 000.1 The at the hospital during the period of the study. There population is rural and poor, with adult unemployment was evidence of an increase in average age at death at 60%. Five per cent of households have piped water of women between 2005 and 2006, suggesting a posi- and 3.6% of households are supplied with electricity. tive impact of the ARV roll-out. The present analysis Government health care in the Ingwavuma sub-district, investigated the continuing impact of HIV/AIDS on in which Mosvold Hospital is situated, is provided by mortality and life expectancy and observed trends over the hospital, 10 residential clinics and 3 mobile clinic the period during which HIV treatment and prevention teams. The hospital mortuary is the only government of mother-to-child transmission therapy (PMTCT) were mortuary serving the Mosvold sub-district. Most deaths introduced. occurring in the sub-district, both within and outside the hospital, are certified by medical staff. The use of nevirapine for PMTCT was commenced in 2002. Dual PMTCT, adding zidovudine to nevirapine, Antiretroviral drugs (ARVs) were first prescribed in Sep- was started in April 2008. tember 2004 as part of the national antiretroviral roll- out programme. Table I shows the total number of pa- Methods tients started on ARVs from 2004 to 2008. The number of females started in each year was greater than the Data from counterfoils of form 83/BI-1663, Notification/ number of males, and from the beginning of the roll-out Register of Death/Stillbirths (Republic of South Africa, at least 11% of the patients enrolled were children. Department of Home Affairs), completed at Mosvold Hospital from 1 January 2003 to 31 December 2008, were entered into a database (Microsoft Access).

36 APRIL 2010 THE SOUTHERN AFRICAN JOURNAL OF HIV MEDICINE Ethical considerations Most deaths between the ages of 20 and 54 years are The publication of statistics on the causes of death cer- due to AIDS-related causes. tified at Mosvold Hospital was approved by the Mos- vold Hospital Ethical Committee. Table II shows the average age at death by year for males and females (>9 years) between 2003 and 2008 accord- Results ing to AIDS-related and non-AIDS-related causes. Aver- age age at death for females declined between 2003 and Figs 1 and 2 show age at death in males and females 2005, and appeared to increase again from 2005 and according to HIV-related and non-HIV-related causes. 2007. The pattern for male deaths is less marked.

TABLE I. PATIENTS STARTED ON ANTIRETROVIRAL DRUGS AT MOSVOLD HOSPITAL, INGWAVUMA, KWAZULU-NATAL, FROM SEPTEMBER 2004

Males Females total patients Cumulative Children <15 yrs Cumulative number % children Year (each year) (each year) started per year total started per year of children <15 yrs <15 yrs 2004 47 78 125 125 14 14 11.2 2005 198 324 522 647 62 76 11.7 2006 367 518 885 1 532 150 226 14.8 2007 397 691 1 088 2 620 170 396 15.1 2008 406 746 1 152 3 772 139 535 14.2

Fig. 1. Total male deaths 2003 - 2008 certified at Mosvold Hospital, Ingwavuma, northern KwaZulu-Natal, grouped according to HIV/AIDS-related and non-HIV/AIDS-related causes of death.

deaths

Fig. 2. Total female deaths 2003 - 2008 certified at Mosvold Hospital, Ingwavuma, northern KwaZulu-Natal, grouped according to HIV/AIDS-related and non-HIV/AIDS-related causes of death. THE SOUTHERN AFRICAN JOURNAL OF HIV MEDICINE APRIL 2010 37 Fig. 3 shows the trends in number of deaths certified for men aged 30 - 34 years, which is consistent with at Mosvold Hospital for the 15 - 59-year age group the mortality patterns found in this study of a popula- and deaths attributable to AIDS-related causes. There tion in the same district. is a 39% reduction in all-cause mortality and a 38% reduction in AIDS-related deaths in women and a 24% In a report by Statistics South Africa entitled ‘Mortality reduction in all-cause mortality in men with a 29% and causes of death in South Africa, 2006’,5 the pro- reduction in deaths attributable to AIDS. portion of deaths according to age group had a similar pattern to that found in this study, with peaks in the Discussion under-5, 30 - 34 and, for females, 75 - 79-year age groups. Deaths in the 15 - 59-year age group increased The National HIV and Syphilis Survey South Africa in between 2002 and 2006, but with a decreasing an- 20073 estimated the antenatal HIV prevalence for Um- nual increase between 2005 and 2006 compared with khanyakude District to be 39.8%, an increase on the previous years. 2006 estimate of 36.3% and higher than the national estimate of 28%. High mortality from HIV/AIDS is con- The pattern of mortality according to age at death and sistent with these estimates. cause of death in this study shows that HIV/AIDS is a leading cause of mortality in persons between the ages In a survey of HIV infection prevalence in the south- of 15 and 59, as well as causing substantial mortality in ern part of Umkhanyakude District, near , the under-5 age group. However, the decline in deaths Tanser et al.4 found that HIV prevalence peaked at 51% in the 15 - 59 age group after 2005, mostly AIDS re- in women in the 25 - 29-year age group and at 44% lated, combined with increased age at death since the

TABLE II. AVERAGE AGE OF DEATH IN PERSONS AGED >9 YEARS - DEATHS CERTIFIED AT MOSVOLD HOSPITAL, INGWAVUMA, KWAZULU-NATAL, JANUARY 2003 - DECEMBER 2008

Males Females Year Average age at death 95% CI Average age at death 95% CI 2003 All causes 47.5 45.8 - 49.2 48.5 46.6 - 50.4 HIV/AIDS 39.4 38.0 - 40.8 35.1 33.7 - 36.5 2004 All causes 45.6 43.9 - 47.3 45.3 43.5 - 47.1 HIV/AIDS 38.8 37.5 - 40.1 36.2 34.9 - 37.5 2005 All causes 45.9 44.2 - 47.6 44.1 42.4 - 45.7 HIV/AIDS 37.9 36.5 - 39.3 35.9 34.6 - 37.2 2006 All causes 45.6 43.8 - 47.3 47.7 45.8 - 49.6 HIV/AIDS 38.7 37.2 - 40.2 35.4 33.9 - 36.8 2007 All causes 47.54 45.7 - 49.4 50.01 48.0 - 52.0 HIV/AIDS 39.62 38.1 - 41.2 37.17 35.5 - 38.8 2008 All causes 47.31 45.3 - 49.3 49.06 46.8 - 51.3 HIV/AIDS 40.62 38.7 - 42.6 35.1 33.4 - 36.8 CI = confidence interval.

Fig. 3. Deaths per year in the age group 15 - 59 years certified at Mosvold Hospital, Ingwavuma, KwaZulu-Natal.

38 APRIL 2010 THE SOUTHERN AFRICAN JOURNAL OF HIV MEDICINE start of the ARV roll-out, suggests that roll-out may be 2. Vaughan Williams CH. Analysis of impact of HIV/AIDS on deaths certified at Mosvold Hospital, Ingwavuma, northern KwaZulu-Natal from 2003 - 2006. South reducing mortality from AIDS-related illness in the 15 - African Journal of Family Practice 2007; 49(5): 16a-16e. http://www.safpj.co.za/ index.php/safpj/article/view/628/756 (accessed 15 June 2009). 59-year age group. The greater impact on female mor- 3. National Department of Health, South Africa. 2008. The National HIV and Syphilis tality compared with male mortality may be explained Survey South Africa 2007. http://www.doh.gov.za/docs/reports/2007/antenatal/ antenatal_report.pdf (accessed 14 June 2009). by the greater number of females compared with males 4. Tanser F, Hosegood V, Bärnighausen T, et al. Cohort profile: Africa Centre enrolled onto ARV treatment. Demographic Information System (ACDIS) and population-based HIV survey. Int J Epidemiol 2008; 37: 956-962. http://www.pubmedcentral.nih.gov/articlerender. fcgi?artid=2557060 (accessed 15 June 2009). REFERENCES 5. Statistics South Africa. 2008. Statistical Release P0309.3: Mortality and causes of 1. Zondi T, Ngomane N. Health and health care systems situational analysis. In: death in South Africa, 2006 Findings from death notification. http://www.statssa. Umkhanyakude Health District Situational Analysis. Braamfontein: Health Systems gov.za/publications/statsdownload.asp?PPN=P0309.3&SCH=4254 (accessed 14 Trust, 2002: 10. June 2009).

THE SOUTHERN AFRICAN JOURNAL OF HIV MEDICINE APRIL 2010 39